AI Art Generator

This script generates AI-based artwork based on a text prompt. It uses various models from both local sources and Hugging Face's model hub. The script can generate a collage of all created images, and also creates an HTML page to display all the images.

Requirements

• PC with an NVIDIA GPU (tested on a Dell Precision 5570 laptop with integrated RTX A1000, also tested with a GeForce RTX 5070ti)
• NVIDIA CUDA drivers installed, see https://developer.nvidia.com/cuda-downloads
• Python 3.x

Text-to-Image Models

By default the script is configured to automatically download and use several models hosted by Hugging Face. The user has the option to manually download and use .safetensors models that can be found on various websites.

Models included by default include:

• plasmo/vox2
• dreamlike-art/dreamlike-photoreal-2.0
• diffusersfan/fking_civitai
• Lykon/DreamShaper
• prompthero/openjourney
• nitrosocke/Ghibli-Diffusion
• SG161222/Realistic_Vision_V1.4
• gsdf/Counterfeit-V2.5
• CompVis/stable-diffusion-v1-4
• runwayml/stable-diffusion-v1-5
• stabilityai/stable-diffusion-2-1

Example Output (index.html)

Installation

First download the repo from GitHub by running this command in your shell:

git clone https://github.com/matt-dreyer/ai_art_generator.git

Create a virtual environment to install into (pretty much required for Linux).

cd ai_art_generator
python3 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip

To install the required libraries, navigate to the directory where the requirements.txt is located and run this command in your shell:

pip install -r requirements.txt

Usage

python art_generator.py [--no-watermark] [--no-collage]

• --no-watermark : Use this flag if you want to disable watermarking on the generated images.
• --no-collage : Use this flag if you want to disable the creation of a collage from the generated images.